HAL Id: hal-01606790
https://hal.archives-ouvertes.fr/hal-01606790
Submitted on 19 Nov 2019
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of
sci-entific research documents, whether they are
pub-lished or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est
destinée au dépôt et à la diffusion de documents
scientifiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.
Distributed under a Creative Commons Attribution - ShareAlike| 4.0 International
License
Maximilian Larter, Sebastian Pfautsch, Jean-Christophe Domec, Santiago
Trueba, Nathalie Nagalingum, Sylvain Delzon
To cite this version:
Maximilian Larter, Sebastian Pfautsch, Jean-Christophe Domec, Santiago Trueba, Nathalie
Na-galingum, et al.. Aridity drove the evolution of extreme embolism resistance and the radiation
of conifer genus Callitris. New Phytologist, Wiley, 2017, 215 (1), pp.97-112. �10.1111/nph.14545�.
�hal-01606790�
Aridity drove the evolution of extreme embolism resistance and
the radiation of conifer genus Callitris
Maximilian Larter
1, Sebastian Pfautsch
2, Jean-Christophe Domec
3,4, Santiago Trueba
5,6, Nathalie Nagalingum
7and
Sylvain Delzon
11BIOGECO, INRA, Univ. Bordeaux, Pessac 33610, France;2Hawkesbury Institute for the Environment, Western Sydney University, Locked Bag 1797, Penrith, NSW 2751, Australia; 3Bordeaux Sciences AGRO, UMR 1391 ISPA INRA, 1 Cours du General de Gaulle, Gradignan Cedex 33175, France;4
Nicholas School of the Environment, Duke University, Durham, NC 27708, USA;5Department of Ecology and Evolutionary Biology, University of California, Los Angeles, UCLA, 621 Charles E. Young Dr. South, Los Angeles, CA 90095, USA;6IRD, UMR AMAP, BPA5, Noumea 98800, New Caledonia;7National Herbarium of New South Wales, Royal Botanic Gardens & Domain Trust, Mrs Macquaries Rd, Sydney, NSW 2000, Australia
Author for correspondence: Maximilian Larter Tel: +33 6 79 70 92 75 Email: maximilien.larter@gmail.com Received: 30 December 2016 Accepted: 26 February 2017 New Phytologist (2017)215: 97–112 doi: 10.1111/nph.14545
Key words: climate change, diversification, drought, ecophysiology, embolism resistance, evolution, gymnosperms, xylem.
Summary
Xylem vulnerability to embolism is emerging as a major factor in drought-induced tree mor-tality events across the globe. However, we lack understanding of how and to what extent cli-mate has shaped vascular properties or functions. We investigated the evolution of xylem hydraulic function and diversification patterns in Australia’s most successful gymnosperm clade, Callitris, the world’s most drought-resistant conifers.
For all 23 species in this group, we measured embolism resistance (P50), xylem specific
hydraulic conductivity (Ks), wood density, and tracheary element size from natural
popula-tions. We investigated whether hydraulic traits variation linked with climate and the diversifi-cation of this clade using a time-calibrated phylogeny.
Embolism resistance varied widely across the Callitris clade (P50:3.8 to 18.8 MPa), and
was significantly related to water scarcity, as was tracheid diameter. We found no evidence of a safety-efficiency tradeoff; Ksand wood density were not related to rainfall. Callitris
diversifi-cation coincides with the onset of aridity in Australia since the early Oligocene.
Our results highlight the evolutionary lability of xylem traits with climate, and the leading role of aridity in the diversification of conifers. The uncoupling of safety from other xylem functions allowed Callitris to evolve extreme embolism resistance and diversify into xeric environments.
Introduction
In the light of current climate change, the role of environmental fluctuation in driving biodiversity and its distribution is of partic-ular interest (Thuiller et al., 2005, 2011). It is now clear that rapid and pronounced changes in environmental conditions can: reduce climatic niche availability, potentially driving populations or species to extinction (Mittelbach et al., 2007; Eiserhardt et al., 2015); open up new suitable niche space allowing range expan-sion; drive adaptation; and even lead to the emergence of new species (Pennington et al., 2004; Hua & Wiens, 2013; Koenen et al., 2015; Qvarnstr€om et al., 2016). For example, the transition to a colder and drier climate in the mid-to-late Miocene (c. 13.8– 5.3 million yr ago (Ma)) is associated with vegetation changes in the Cape Flora of South Africa, leading to a major increase in diversification in Aizoaceae (Dupont et al., 2011). Over the same period, the South American orchid genus Hoffmannseggella radi-ated into more open habitats with the cooling climate in eastern Brazil (Antonelli et al., 2010). The emergence of the sclerophyll biomes in Australia 25 Ma has also been associated with adaptive
radiations, with evidence from the fossil record (Carpenter et al., 1994; Hill, 2004) and using phylogenetics, for example, in Banksia and eucalypts (Crisp et al., 2004). Although we do not fully understand the rise of the angiosperms (Augusto et al., 2014), climate shifts at the end of the Eocene c. 34 Ma could also explain the decline of some gymnosperm lineages (Crisp & Cook, 2011).
Although some species have shown great capacity for adapta-tion (Meyers & Bull, 2002) and shifts in distribuadapta-tion (Davis & Shaw, 2001), reports of potential habitat contraction and increas-ing extinction risk of tree species as a consequence of today’s rapid climate change are increasing (e.g. Eiserhardt et al., 2015; Corlett & Westcott, 2013; Sax et al., 2013). It thus becomes apparent that in order to better predict climate change-induced effects on diversity and habitat range of trees, it is necessary to identify adaptive traits that could offer resistance to extinction and have the potential to respond to natural selection during rapid climate change.
For vascular plants, and tall trees in particular, providing a continuous stream of water to their photosynthetic aerial shoots
is one of their most vital needs and greatest challenges, especially as climate change will increase frequency and severity of droughts and heat waves (Stocker et al., 2013; Pfautsch, 2016). Water transport in plants is driven by the process of transpiration at the leaf–atmosphere interface, which generates a water-potential gra-dient throughout the plant. Formalized by the cohesion-tension theory (Dixon, 1914; Tyree & Zimmermann, 2002), water transport is possible thanks to the strong hydrogen bonds between water molecules that allow liquid water to remain in a metastable state while at negative pressure, i.e. under tension. The main drawback of this remarkable function is the risk of xylem embolism, or breakage of the water column because of cav-itation, which becomes ever more likely as evaporative demand increases, and as soil water potential declines during drought. After embolism, the air-filled conduits restrict water transport, and high amounts of xylem embolism can lead to plant death (Brodribb & Cochard, 2009; Urli et al., 2013). The breakdown of the vascular system through embolism is thought to be involved in multiple mass mortality events during and after droughts in forests across the globe in recent decades (Anderegg et al., 2012, 2016). Various adaptive mechanisms protect plants from these potentially catastrophic events, such as limiting water loss through transpiration and the drop in xylem pressure by clos-ing stomata early durclos-ing drought (Delzon & Cochard, 2014), or xylem adaptations that limit the formation of air bubbles and the irreversible spread of embolism (Schuldt et al., 2015; L€ubbe et al., 2016). Recent advances have revealed that cavitation events occur at air–sap interfaces within pores between functional and embolized conduits (Choat et al., 2008; Lens et al., 2013). In conifers, bordered pits contain a valve-like structure called the torus which is deflected to block the pit aperture, protecting functional tracheids from air entry (Bailey, 1916). While this provides conifers with increased protection from embolism com-pared with angiosperms (Maherali et al., 2004; Pittermann et al., 2005; Choat et al., 2012), embolism still occurs as a result of imperfect sealing of the pit aperture by the torus (Bouche et al., 2014).
Plants need to balance embolism resistance (xylem safety) with optimizing rates of water transport (xylem efficiency) to their photosynthetic organs (Tyree & Zimmermann, 2002; Hacke et al., 2006). The advantages of xylem safety are evident (i.e. increased survival during drought), whereas xylem efficiency allows increased carbon allocation to leaves relative to sapwood area, for example, allowing rapid growth and optimizing photo-synthesis in competitive settings (Santiago et al., 2004; Poorter et al., 2010). The link between safety and efficiency is indirect, as wider conductive elements transport water efficiently but are weaker under the mechanical stress imposed by negative xylem pressure in drought-tolerant species (Sperry et al., 2008). Thicker cell walls could influence pit morphology, notably the pit mem-brane, thereby increasing xylem embolism resistance (Tyree & Sperry, 1989; Li et al., 2016). The final compromise reached between safety and efficiency is strongly determined by species ecology and phylogeny. For example, species growing in compet-itive wet environments are more likely to maximize efficiency over safety (Sperry et al., 2006; Choat et al., 2012). In any case,
there is little evidence of a universal and strong relationship between safety and efficiency at a higher ecological and taxo-nomic scale. This tradeoff is only suggested by the absence of species with a xylem that combines both embolism resistance and transport efficiency (Gleason et al., 2016). However, there are many ‘incompetent’ species, i.e. with vasculature that is both vul-nerable to embolism and inefficient for water transport. The impact of water availability on water transport efficiency seems, however, to be noticeable at more restricted evolutionary scales. For instance, xeric Eucalyptus species evolved narrower vessels and denser wood than species in more mesic environments, reflecting a tradeoff of hydraulic traits in response to climate across Australia (Pfautsch et al., 2016). Similarly, tracheid lumen diameter tracks rainfall across Callitris columellaris populations (Bowman et al., 2011).
At broad evolutionary scales there is substantial variation in the xylem pressure inducing 50% loss of hydraulic conductance, termed P50 (Maherali et al., 2004; Pittermann et al., 2012;
Bouche et al., 2014). On the other hand, at finer infrageneric and intraspecific scales there is generally low variability (Delzon et al., 2010) and little genetic differentiation for xylem embolism resis-tance (Lamy et al., 2011; Saenz-Romero et al., 2013; but see David-Schwartz et al., 2016; Hajek et al., 2016). However, no studies have examined embolism resistance among closely related species over a wide climatic gradient. Callitris is a conifer genus of shrubs and trees that underwent an ecological radiation with the emergence of dry environments in Australia over the last 30 million yr (Pittermann et al., 2012). It is the largest genus of Cupressaceae in the southern hemisphere, with 15–20 species recognized by different authors (Hill & Brodribb, 1999; Farjon, 2005; Eckenwalder, 2009; Piggin & Bruhl, 2010). This clade dis-plays a marked xeric affinity, despite spanning a huge rainfall gra-dient across Australia and New Caledonia, i.e. from c. 200 to over 2000 mm yr1. With some of the most drought-tolerant species currently known (Brodribb et al., 2010; Larter et al., 2015), this clade presents an ideal group of species to investigate plant ecophysiology and the development of drought tolerance from an evolutionary perspective.
Our main objective was to investigate the evolution of the vas-cular system of all members of the Callitris clade. First, we con-structed an extensive physiological dataset of xylem embolism resistance, xylem specific hydraulic conductivity and xylem anatomical traits. We then obtained species occurrence data and tested whether variation in hydraulic traits was linked to species climatic preferences and the diversification of this clade, estab-lished with a time-calibrated phylogeny based on DNA sequence data from multiple loci. We hypothesized that, owing to the large aridity gradient occupied by this clade, resistance to embolism should vary widely and track climatic conditions. Given the genus-level phylogenetic scale and extensive climatic gradient, our second hypothesis was that a tradeoff existed between safety and hydraulic efficiency, as a result of large selective pressures from competition and climatic stress driving hydraulic traits. Specifically, we expected the most xeric species to have a resistant, less efficient xylem with many narrow tracheids, while species from more mesic regions would have wider tracheids, favoring
xylem efficiency but reducing safety from embolism. Our third hypothesis was that if hydraulic traits in this group evolved rapidly in response to climate, then phylogeny would have little impact on traits and tradeoffs. In relation to this hypothesis, we expected that the distribution of members of the Callitris clade reflected an ecological radiation in response to increasing aridity in the Australian region since the end of the Eocene c. 34 Ma.
Materials and Methods
Collection of plant materials
For the purpose of this study, we collected samples from wild populations for Actinostrobus acuminatus, Actinostrobus arenarius, Actinostrobus pyramidalis, Callitris canescens, Callitris drummondii, Callitris endlicheri, Callitris glaucophylla, Callitris gracilis, Callitris muelleri, Callitris neocaledonica, Callitris roei, Callitris sulcata, Callitris verrucosa, and Neocallitropsis pancheri and from botanical gardens for Callitris baileyi, Callitris endlicheri, Callitris macleayana, and Callitris monticola (Table 1; Fig. 1). We also included data from previous studies for Callitris tuberculata, C. columellaris, C. gracilis, Callitris preissii, Callitris rhomboidea and A. pyramidalis (Brodribb et al., 2010; Delzon et al., 2010; Larter et al., 2015), and pooled data from all sources for analysis. The close relationship between the genera Callitris s.s., Neocallitropsis and Actinostrobus (Pye et al., 2003; Piggin & Bruhl, 2010) led us to treat them as a single Callitris clade (or Callitris s.l.). We collected straight branches c. 40 cm long and 1 cm in diameter, using hand-held secateurs or telescopic branch-cutters from sun-exposed parts of the crown, although we expect little variation in xylem embolism resis-tance across organs in conifers (Bouche et al., 2016). We sampled three to 10 individuals depending on the availability of healthy adult trees. We immediately removed all leaves on the branches, and carefully wrapped the samples (in wet paper towels, plastic bags and packing tape), to avoid any water loss during courier transport to analytical facilities in Bordeaux, France. They were then shipped as quickly as possible, refrigerated and maintained in a humid environment until hydraulic measurements were conducted. Species climate
The climate experienced by Callitris species varies widely from hot, semiarid Mediterranean climate in southwestern and cen-tral Auscen-tralia to wet tropical or monsoonal climates in North-ern Australia, Queensland and New Caledonia (Fig. 1). In order to describe each species’ climate, we downloaded species occurrence points from the Global Biodiversity Information Facility (GBIF; Booth, 2014). After excluding data with pos-sible species misidentification or GPS coordinate errors, we extracted climate information for each point using climate data from the WorldClim database (Hijmans et al., 2005) using QGIS software (Quantum, 2011). We also obtained potential evapotranspiration (P-ET) and aridity index (AI) from the CGIAR-CSI database (Trabucco & Zomer, 2009). We then extracted the mean and median values of the distri-butions of each climatic variable for each species.
Physiological and anatomical traits
All hydraulic measurements were done with the standard proto-col developed at the University of Bordeaux using the ‘Cavitron’ technique (Cochard et al., 2005), which uses centrifugal force to induce a negative pressure in a spinning sample. The loss of hydraulic conductance as a result of embolism is monitored while incrementally decreasing the negative pressure experienced in the xylem. We used a thicker, reinforced rotor specifically designed to safely reach high speeds (26 000 g, or 13 000 rpm with a stan-dard 27.5 cm rotor), inducing xylem pressure below16 MPa. This technique allowed us to construct vulnerability curves and simultaneously estimate xylem specific hydraulic conductivity Ks
(kg m1MPa1s1), derived from the maximum conductance measured at high xylem pressure divided by sample length and sapwood area. A sigmoid model was fitted to each individual vul-nerability curve (207 in total) from which we obtained the P50
parameter (MPa), which is the xylem pressure inducing 50% loss of conductance (Pammenter & Vander Willigen, 1998). Values were then averaged across all samples for each species.
Following xylem embolism resistance measurements, a 1–2 cm segment was excised from the basal part of the samples in order to conduct xylem anatomy measurements. Several transverse sec-tions per individual were cut using a sliding microtome, stained using safranine at 1%, and examined using a light microscope (DM2500M; Leica Microsystems, Wetzlar, Germany). We selected three individuals per species and took five digital images per individual, which were then analyzed using ImageJ (NIH, Bethesda, MD, USA). Magnification was9400 for all samples except for N. pancheri, C. sulcata and C. neocaledonica, for which we used9200 magnification to increase the total number of tra-cheids visible on each image. Sample preparation and photogra-phy were conducted at either the University of Bordeaux or Western Sydney University (Hawkesbury Institute for the Envi-ronment), using the same protocol. Images were manually edited to remove imperfections before automatic image analysis. Tracheids with an incomplete lumen (i.e. placed on the edge of the images) were excluded from the analysis. Overall, 80 individ-ual samples (three to five per species) and a total of 400 images were analyzed, resulting in a total of 33 652 tracheids measured. We extracted from each image the area of each tracheid lumen, total image area and number of analyzed tracheids, and pooled the data by individual for analysis. From the total analyzed area, total tracheid number, and area of each tracheid, we derived for each individual mean, minimum and maximum tracheid diame-ter, the average number of tracheid per unit area (TF; n mm2) and the ratio of tracheid lumen area divided by total image area (void to wood ratio; %).
According to the Hagen–Poiseuille law, hydraulic conductivity of a conduit varies according to the fourth power of its diameter, which means that larger tracheids contribute disproportionately more to the overall hydraulic conductance (Tyree & Zimmermann, 2002). To use a hydraulically meaningful measure of tracheid diameter, we calculated the hydraulically weighted hydraulic diameter (Dh; Tyree & Zimmermann, 2002), defined
Tabl e 1 Samp ling deta ils for all specie s in this study (natu ral populat ions and botanic garden s location s) and media n val ues for aridity index (AI ), mean ann ual tempe rature (M AT) and mean annual precip itatio n (MAP) of each specie s’ distribu tion Spe cies Author ity N Nat ural popu lation locat ion Lati tude Long itude Elevatio n (m)
Botanic garden location
Lat itude Longitude Elevati on (m ) A I MAT MAP Refere nce Acti nostrob us acum inatus Parl. 10 Wes tern Aust ralia 30.4 94 28 115.42 2 5 3 1 9 5 0.39 18.5 576 Acti nostrob us aren arius C.A.Gard ner 11 Wes tern Aust ralia 30.8 61 39 116.72 1 6 7 3 3 8 Mt An nan BG, N SW, Aus. 34 .06984 150.764 94 155 0.26 19.15 408.5 Acti nostrob us pyra midal is Miq. 12 Wes tern Aust ralia 30.6 27 31 115.94 1 3 6 2 0 4 0.48 18.2 698 1 Ca llitris baileyi C.T.W hite 3 M t A n nan BG, N SW, Aus. 34 .06984 150.764 94 155 0.59 18.1 878 Ca llitris cane scens 1 (Parl.) S.T.Bla ke 10 Wes tern Aust ralia 33.1 10 78 118.35 4 2 8 3 0 5 Mt An nan BG, N SW, Aus. 34 .06984 150.764 94 155 0.28 16.5 372 10 Sout h Aust ralia 34.9 99 25 139.55 5 3 3 7 7 Ca llitris colume llar is F.Mue ll. 3 0.22 19.6 367 2 Ca llitris dru mmond ii (Parl.) Ben th. & Hook. f. ex F.Mue ll. 13 Wes tern Aust ralia 33.6 78 81 120.18 5 6 1 1 3 4 0.37 16.6 495 Ca llitris endlich eri F.M.B ailey 9 New South Wales 33.1 37 00 150.07 4 6 1 6 1 2 0.49 15.7 696 Ca llitris glauco phyl la J.Thomp s. & L.A.S.J ohns on 13 New South Wales 32.4 34 81 149.84 6 1 9 7 0 7 Mt An nan BG, N SW, Aus. 34 .06984 150.764 94 155 0.32 17.2 459 Ca llitris gracili s R.T.Bak er 8 Sout h Aust ralia 34.0 72 93 140.79 8 8 8 5 7 0.32 15.8 421 2 Ca llitris int ratrop ica R.T.Bak er & H.G.Sm. 5 Royal BG Sydne y, NSW, Aus. 33 .86557 151.214 15 30 0.69 26.8 1232.5 Ca llitris macle ayana (F.Muel l.) F. M uell. 3 M t A n nan BG, N SW, Aus. 34 .06984 150.764 94 155 1.32 17.7 1690 Ca llitris mon ticola J.Garden 7 M t T o mah BG, N SW, Aus. 33 .53942 150.419 62 950 0.94 14.25 1220 Ca llitris mueller i (Parl.) Ben th. & Hook. f. ex F.Mue ll. 8 New South Wales 33.7 32 42 150.37 1 9 2 8 1 5 1.07 13.9 1233 Ca llitris neocal edon ica D €umme r 11 New Ca ledonia 21.8 82 50 166.41 6 3 9 1391 2.03 18.3 2320 Ca llitris oblong a Rich. & A.Ri ch 4 Royal BG Kew, UK 51 .43539 0.751 73 53 0.8 12.8 918 Ca llitris preissi i Miq. 3 0.26 17 363 2
Dh¼ P D4 N 1=4 Eqn 1 where N is the total number of tracheids and Dh is the mean
diameter required to achieve the same overall conductance with the same number of conductive elements. We then calculated the theoretical conductance (Kth; kg m1MPa1s1), which is the
theoretical rate of flow through a cylindrical pipe according to the Hagen–Poiseuille law:
Kth ¼ D 4 hpq
128 g TF Eqn 2
where g is the viscosity of water at 20°C (1.002 9 109MPa s)
and q is the density of water at 20°C (998.2 kg m3). For
appro-priate units, we transformed Dhto m (9 106) and TF to m2
(9 105).
Xylem water flow efficiency represents the joint conductivity of both tracheid lumens (i.e. Kth) and bordered pits (Hacke et al.,
2005). We calculated pit conductivity Kpit (kg m1MPa1s1)
as follows: Kpit ¼ 1 Ks 1 Kth 1 Eqn 3 We macerated sapwood of 14 Callitris species and measured length (lm) and counted bordered pits (hereafter pitting) of 700 individual tracheids (50 per species), for a subset of species resentative of the range of climate encountered by Callitris. To pre-pare maceration slides (after Franklin, 1945), splinters of outer sapwood were incubated for 12 h at 60°C in plastic vials (2 ml Eppendorf) that contained acetic acid and hydrogen peroxide (50 : 50). Following the incubation, splinters were repeatedly washed and dehydrated in ethanol (50–70–96%) with water before staining with safranin–ethanol solution. After excess stain was removed, samples were mounted on glass slides, and cells were imaged using a light microscope and a digital camera (Leica DFC 500); ImageJ was again used for image analysis.
Wood density was estimated using X-ray imagery (Polge, 1966) for three individuals per species on transversal sections of c. 1 mm thickness. We used a silicone scale of known densities for calibration, and analyzed the images using Windendro (Guay et al., 1992) to obtain two radial density profiles per section, from which we estimated mean wood density (g cm3).
Phylogenetic reconstruction
We built on the increased availability of molecular sequences to produce a new phylogeny of Callitris, Actinostrobus and Neocallitropsis. We took a broad view of species delimitations, i.e. treating C. columellaris, Callitris intratropica, and C. glaucophylla separately, as well as C. tuberculata, C. preissii, C. verrucosa and C. gracilis. Various treatments of these as subspecies and/or synonyms are available in the literature (Hill, 1998; Hill & Brodribb, 1999; Farjon, 2005, 2010; Eckenwalder, 2009). We
Tabl e 1 Contin ued. Spe cies Author ity N Nat ural popu lation locat ion Lati tude Long itude Elevatio n (m)
Botanic garden location
Lat itude Longitude Elevati on (m ) A I MAT MAP Refere nce 2 Ca llitris rho mboidea R.Br. ex Rich. & A.Rich. 8 0.83 14.6 924 1, 2 Ca llitris roei (Endl.) F.Mue ll. 9 Wes tern Aust ralia 33.1 08 06 118.33 4 3 9 3 4 6 0.29 16.2 387 Ca llitris sulcat a (Parl.) Sc hltr. 9 New Ca ledonia 22.1 33 65 166.48 8 8 7 1 6 0.98 22.5 1252 Ca llitris tub erculata R.Br. ex R.T.Bak er & H.G.Sm. 11 Wes tern Aust ralia 33.1 10 78 118.35 4 2 8 3 0 5 Mt An nan BG, N SW, Aus. 34 .06984 150.764 94 155 0.21 17.3 304 3 Ca llitris verruco sa (A.Cun n. ex Endl .) R.Br. ex Mirb . 9 Sout h Aust ralia 35.0 18 28 139.46 5 6 1 7 1 0.25 16.3 345 Neoca llitropsi s panch eri (Carri ere) de Laub. 18 New Ca ledonia 22.2 31 94444 166.86 0 5556 24 4 1.79 21.5 2162 1 C. cane scens was samp led from two wild popu lations , in Weste rn Austral ia and South Austr alia, as well as on e indiv idual from Mt Anna n BG, NSW. 2 Refere nces for data ta ken fro m prev ious st udies: 1, D elzon et al. (2010 ); 2, Bro dribb et al. (2010 ); 3, Larter et al. (20 15).
used the PHLAWD pipeline (Smith et al., 2009) to obtain sequences from GenBank (Benson et al., 2011) with the aim of constructing a large multilocus dataset. We used a range of loci from plastid (rbcL, matK, trnL, psbB, petB) and nuclear DNA (ITS, needly, leafy). To complete the dataset, we used unpub-lished sequences (rbcL and matK) for the species C. roei, C. monticola, C. baileyi, and C. columellaris (N. Nagalingum, pers. comm.).
Alignments from PHLAWD were manually trimmed and checked, and we selected the most appropriate DNA substitution models using maximum likelihood implemented in MEGA (Tamura et al., 2011) (see Supporting Information Table S1 for details). To check for incongruence between the chloroplastic and nuclear datasets, we first analyzed both matrices separately using maximum likelihood with RAXML (v.8.0.20) and bayesian inference implemented in BEAST v.1.8.2 (Drummond & Ram-baut, 2007) (Fig. S1). Finally, we combined the data to construct a time-calibrated phylogeny for the subfamily Callitroidae (Cupressaceae). We used the recognized position of Papuacedrus as sister to all other species in this group to root the tree (Leslie et al., 2012; Mao et al., 2012). We set a minimum constraint on the age of crown Callitroidae (Leslie et al., 2012; Mao et al., 2012) by using the Patagonian fossil Papuacedrus prechilensis (Wilf et al., 2009), based on an ovulate cone from the early to mid-Eocene (51.9–47.5 Ma). The maximum age for the appear-ance of crown Callitroidae was conservatively set by evidence that crown Cupressaceae (i.e. subfamilies Callitroidae and Cupres-soidae) was present by 99.6 Ma, using leafy shoots and cones of the fossil Widdringtonia americana from the Cenomanian (McIver, 2001). Although this fossil is probably too old to belong to the extant genus Widdringtonia, it definitely belongs to the
Cupressaceae s.s., that is, excluding the Taxodioid lineages. We used the earliest evidence of Callitris (Paull & Hill, 2010) to con-strain the crown of the Callitris clade. C. leayana from early Oligocene Tasmania has cone scales and leaves in whorls of three, which are easily distinguishable from Fitzroya, the only other extant genus with this arrangement. However, this fossil bears characters of both Callitris and Actinostrobus, and the phyloge-netic relationships between the two genera are still ambiguous (Pye et al., 2003; Piggin & Bruhl, 2010). Finally, we used Fitzroya acutifolia to calibrate the split between Fitzroya and Diselma, based on the high similarity between the fossil leaves and cones from the early Oligocene and extant Fitzroya (Paull & Hill, 2010). We used a uniform prior for the root calibration (49.9–99.6 Ma), and lognormal priors for the crown Callitris clade and the Fitzroya–Diselma split, specified so that the 95% confidence interval (CI) spanned 28.3–48.3 Ma. We ran the chain for 100 million generations, sampling every 10 000. We discarded the first 25% of the chain as burn-in and evaluated convergence using TRACER v.1.6– all effective sample sizes were well above 200.
Statistical analyses
Vulnerability curve fitting was done in SAS 9.4 (SAS Institute Inc., Cary, NC, USA). All remaining data manipulation and analysis were run in R v.3.2.3 (R Core Team, 2015). We per-formed linear regressions on raw data and log-transper-formed data with nearly identical results, so we present only untransformed correlations to simplify interpretation. When significant rela-tionships appeared nonlinear, we fitted a nonlinear model (i.e. y= a9log(x) + b) with nonlinear least-squares (NLS function) in Deserts and xeric shrublands
Mediterranean forests, woodland and scrub Montane grasslands and shrublands Temperate broadleaf and mixed forests Temperate grasslands, savanas and shrublands Tropical and subtropical moist broadleaf forests Tropical and subtropical dry broadleaf forests
Tropical and subtropical grasslands, savannas and shrublands
Actinostrobus arenarius Actinostrobus pyramidalis Actinostrobus acuminatus
Callitris tuberculata Callitris canescens
Callitris roei Callitris drummondii Callitris glaucophylla Callitris endlicheri Callitris muelleri Callitris sulcata Callitris neocaledonica Callitris gracilis Callitris verrucosa Neocallitropsis pancheri
Fig. 1 Map of locations of wild populations sampled for this study. Colors represent terrestrial eco-regions taken from Olson et al. (2001).
R. However, to allow comparisons across bivariate relation-ships, in all cases we report the R2 values from the linear cor-relations. Phylogenetic generalized least squares (PGLS) were run using the CAPER package v.0.5.2 (Orme, 2013). PGLS accounts for the statistical nonindependence in cross-species trait correlations by adjusting the residual error structure using a variance/covariance matrix derived from the phylogeny (Garamszegi, 2014). It explicitly allows for varying amounts of phylogenetic signal in the data by using Pagel’s lambda (Pagel, 1999), thus removing the risk of overcorrecting for phylogeny when shared evolutionary history does not affect trait relation-ships. We used PHYTOOLSv.0.5 (Revell, 2013) to map the evo-lution of P50 onto the time-calibrated phylogeny using a
Brownian motion model of evolution and maximum likeli-hood. We used APE v.4.0 (Paradis et al., 2004) to estimate ancestral states, to plot the number of lineages through time, and to obtain the gamma statistic (Pybus & Harvey, 2000).
Results
Variation in hydraulic traits
Xylem embolism resistance varied widely across the Callitris clade, from P50= 3.8 0.1 MPa in C. neocaledonica to
18.8 0.6 MPa in C. tuberculata (mean SE; Fig. 2; Table S2). P50had a mean of11.7 0.7 MPa across the whole
clade. Xylem specific hydraulic conductivity (Ks) varied
approxi-mately sixfold, from 0.187 0.014 kg m1MPa1s1 in C. oblonga to 1.25 0.05 kg m1MPa1s1 in C. intratropica, with a clade-wide mean of 0.52 0.05 kg m1MPa1s1. Our analysis of xylem anatomy uncovered extensive variation, with some species containing many small, densely packed tracheids (e.g. C. tuberculata, mean diameter= 9.04 0.23 lm, tracheid frequency= 3250 126 n mm2), while others had fewer but wider tracheids (e.g. C. sulcata, 14.14 0.77 lm and
Neocallitropsis pancheri Callitris macleayana Callitris sulcata Callitris preissii Callitris canescens Callitris columellaris Callitris intratropica Callitris glaucophylla Callitris tuberculata Callitris verrucosa Callitris gracilis Callitris endlicheri Callitris rhomboidea Callitris muelleri Actinostrobus acuminatus Actinostrobus pyramidalis Actinostrobus arenarius Callitris drummondii Diselma archeri Widdringtonia nodiflora Libocedrus bidwillii Libocedrus yateensis Pilgerodendron uviferum Papuacedrus papuana Callitris baileyi Callitris oblonga Callitris monticola Callitris roei Fitzroya cupressoides P50 (MPa) –4.1 –18.8 A B C D E 0 1 0 0 3 20 0 5 40 0 7 60 0 9 80 100 P. e n e c o i M Pli. Oligocene Eocene Paleocene Late Cretaceous
Time before present (million yr ago)
1 * * * 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 –10 –20 P50 (MPa) (a) (b)
Fig. 2 Time-tree of Callitroidae and the evolution of resistance to embolism (P50) in the Callitris clade. (a) Phylogeny of Callitroidae. Blue bars at nodes represent 95% highest posterior density from the BEASTanalysis, and numbers represent Bayesian posterior probabilities (not shown if< 0.95). Branches of
the phylogeny are colored according to P50reconstructed using the Brownian motion model (scale bottom left). Stars mark fossil-calibrated nodes. Letters A–E indicate the five subclades within crown Callitris. (b) Bar plot of P50for each species in the phylogeny, colored according to species aridity index (AI). Red, AI< 0.5; green, 0.5 < AI < 1; blue, AI > 1. Callitris neocaledonica is the only missing species: P50= 3.8 MPa, AI = 2.0. Pli., Pliocene; P., Pleistocene.
1607 178 n mm ). Accordingly, theoretical hydraulic con-ductance (Kth) also displayed a wide range of values, with the
xylem of C. sulcata c. 10 times more efficient than that of C. gracilis (13.2 1.7 and 1.15 0.14 kg m1MPa1s1, respectively). Comparatively, variation in wood density was mod-erate yet significant, from 0.57 0.02 g cm3in C. oblonga to a maximum of 0.74 0.04 g cm3in N. pancheri.
Using occurrence data, we found that across the geographical range of the Callitris clade, mean annual precipitation (MAP) varied from c. 300 mm for C. tuberculata in Western Australia to over 2100 mm for N. pancheri in New Caledonia. Similar varia-tion was found for the aridity index (AI), ranging from 0.21 to 2.03 for C. tuberculata and C. neocaledonica, respectively. Mean annual temperature varied from 12.8°C for C. oblonga to 26.8°C for C. intratropica.
We tested the role of water availability in determining species’ hydraulic properties. More negative P50 was strongly related to
increasing aridity across species ranges (Fig. 3a,b; R² = 0.78 and 0.81 for MAP and AI, respectively). This trend was confirmed using PGLS, proving the adaptive role of xylem embolism resis-tance in Callitris. There was nonetheless wide variation in P50for
similar rainfall values. For example, in the most arid environ-ments with MAP < 450 mm, embolism resistance was in the range P50= 12 to 19 MPa. Tracheid dimensions also track
rainfall, with increasing precipitation associated with wider tra-cheids (Fig. 3g,h; R² = 0.45 and 0.48 for MAP and AI, respec-tively). Neither Ks(Fig. 3d–f) nor wood density (Fig. 3j–l) were
linked to MAP in the studied clade. However, these traits were related to mean annual temperature, with warmer climates linked to higher transport efficiency and higher wood density, although R² values were low (Fig. 3f,l; R² = 0.245 and 0.135, respectively). By contrast, we found no evidence for an effect of temperature on P50or tracheid diameter (Fig. 3c,i). Accounting for
phyloge-netic proximity using PGLS did not impact these trait–climate relationships (Table S3). We note that since MAP enters into the calculation of AI, these two variables are strongly correlated.
In our dataset, we found that P50and tracheid diameter were
strongly positively correlated, and P50 and tracheid frequency
were negatively correlated, with resistant species having numerous smaller tracheids compared with species that were more vulnera-ble (Fig. 4a,b). No relationship was detected between P50 and
void-to-wood ratio or wood density (Fig. 4c,d). In turn, Ks
appeared to be disconnected to some extent from tracheid and wood traits, showing only a significant relationship with tracheid frequency (Fig. 4f). The remarkably high conductance measured for C. intratropica must be highlighted, as it is the result of a highly unusual combination of hydraulic traits: narrow Dh
(Fig. 4e), low tracheid frequency (Fig. 4f), and moderate void-to-wood ratio (Fig. 4g) and void-to-wood density (Fig. 4h). Overall, hydraulic safety (P50) and efficiency (Ks) were not closely linked
to wood traits in the Callitris clade, yet resistant species developed smaller, more frequent tracheids. We also investigated the safety– efficiency tradeoff hypothesis, but we found no significant rela-tionship between P50and Ks(Fig. 5). In this line, bordered pit
conductivity was unrelated to P50(Fig. 5c). By contrast,
theoreti-cal maximum hydraulic conductance (Kth) was negatively
correlated with xylem embolism resistance. Resistant species had a theoretically less efficient xylem when only considering tracheid lumen conductivity (Fig. 5b), which is consistent with smaller tracheids being related to embolism resistance (Fig. 4a). By con-trast, bordered pit conductivity (Kpit) was unrelated to P50, and
remained around or below 1 kg m1MPa1s1 across species (Fig. 5c). These relationships were not impacted by phylogenetic history (Table S4). Finally, tracheid length and average number of pits per tracheid were weakly correlated with embolism resis-tance, where more embolism-resistant species had shorter tra-cheids with fewer pits (e.g. C. tuberculata, 1167 35 lm and 38 2 pits) than more vulnerable species (e.g. C. macleayana, 1931 43 lm and 81 3 pits). However, higher pitting was not associated with an increase in overall conductance (Fig. S2). Phylogeny and trait evolution
Our phylogenetic analyses recovered the well-recognized relation-ships within the subfamily Callitroidae, with the Pilgerodendron– Libocedrus clade sister to remaining species. The Diselma-Fitzroya clade is closely allied to Widdringtonia (Fig. 2). Within the Callitris s.l. group containing the paraphyletic Callitris, Actinostrobus and Neocallitropsis, early divergences dated to between 30 and 25 Ma are not well resolved (Fig. 2). We found five separate subclades (see letters in Fig. 2): C. drummondii and C. baileyi (A), the three Actinostrobus species (B), two tropical northeastern clades composed of the New Caledonian species N. pancheri and C. sulcata (B) and C. macleayana (C) and a well-supported core Callitris clade encompassing the remaining Aus-tralian species (E), which is comfortably the most species-rich group within Callitris (14 species).
When examining diversification trends, our phylogenetic anal-ysis indicates an upward shift c. 30 Ma (Fig. 6a), with the rapid successive divergences of the five subclades identified in Fig. 2a. A second acceleration in diversification corresponds with the diver-gences in the core Callitris clade (clade E in Fig. 2a), c. 17 Ma and continuing to the present. Tempo of diversification varied through time, with a c-statistic of 2.36 (P= 0.02), indicating either an early decrease in extinction rates or a pulse of speciation toward the present (Pybus & Harvey, 2000). By mapping the evolution of xylem embolism resistance onto the phylogeny, we showed that several lineages achieved extreme P50 under
15 MPa independently over the last 10 million yr, i.e. A. arenarius, C. canescens, C. columellaris and C. tuberculata (Fig. 2a). Overall, P50was extremely labile, with large differences
between closely related species, for example C. monticola and C. roei (P50= 8.8 and 12.2 MPa, respectively), C. oblonga and
C. canescens (10.9 and 16.9 MPa), N. pancheri and C. sulcata (4.1 and 8.2 MPa) and A. pyramidalis and A. arenarius (11.8 and 15.2 MPa). Related species of the Callitroideae subfamily are relatively vulnerable to xylem embolism, e.g. Papuacedrus papuana (P50= 4.7 MPa) and Pilgerodendron
uviferum (P50= 4.4 MPa). The ancestral P50 reconstruction
inferred that the ancestral state in the Callitroidae subfamily was c.7.2 MPa (95% CI: 1.6 to 12.4 MPa), and that the com-mon ancestor of the Callitris clade was moderately resistant, with
Ks (kg m –1 MPa –1 s –1 ) P50 (MPa) 0 500 1000 1500 2000 2500 –20 –15 –10 –5 0 R² = 0.783 *** C_in C_tu N_pa
Mean annual precipitation (mm)
0 500 1000 1500 2000 2500 (a) (b) (c) (f) Dh (μm) 10 12 14 16 18 20 22 24 26 R² = 0.455 *** C_in N_pa (g) (i) 0.5 0.6 0.7 0.8 R² = −0.042 ns C_in C_tu N_pa W o od density (g cm –3 ) (j) (l) 0 0.5 1 R² = 0.086 ns 1.5 (d)
Aridity index (MAP/P-ET)
0.0 0.5 1.0 1.5 2.0 R² = 0.805 *** C_tu N_pa R² = 0.008 ns 0.0 0.5 1.0 1.5 2.0 R² = 0.476 *** N_pa 10 14 18 22 26 R² = −0.04 ns R² = 0.245 ** C_in
Aridity index (MAP/P-ET)
R² = −0.043 ns
C_tu N_pa
R² = 0.083 ns
Mean annual temperature (°C)
10 14 18 22 26
R² = 0.135 *
Mean annual temperature (°C)
C_in N_pa N_pa C_in N_pa C_tu (e) (h) (k) C_in C_in C_tu C_tu C_tu C_tu
Mean annual precipitation (mm)
C_in C_tu N_pa C_in C_tu N_pa C_in C_tu N_pa C_in
Fig. 3 Relationships between xylem hydraulic traits and climate in Callitris. (a–l) Embolism resistance (P50; a–c), xylem specific hydraulic conductivity (Ks; d–f), hydraulically weighted tracheid diameter (Dh; g–i) and wood density (j–l) in relation to water availability, as measured by mean annual precipitation (MAP, left), aridity index (AI, centre) and mean annual temperature (right). Linear regression lines are shown, with the adjusted R². Statistical significance is indicated by asterisks:***, P < 0.001; **, P < 0.01; *, P < 0.05; ns, nonsignificance at a = 0.05. Dashed curves in (a) and (b) represent the best fit nonlinear model. C_tu, Callitris tuberculate; C_in, C. intratropica; N_pa, Neocallitropsis pancheri; P-ET, potential evapotranspiration. Red, AI< 0.5; green, 0.5< AI < 1; blue, AI > 1. Error bars represent SE.
P50~10.4 MPa (95% CI: 8 to 12.9 MPa). Details for the
reconstructed P50 values at each node of the phylogeny can be
found in Fig. S3.
Discussion
Evolution of P50and the adaptive radiation of Callitris
Our results provide strong evidence that the evolution of extreme xylem embolism resistance accompanied the radiation of the Callitris clade and was driven by the aridification of the Australian continent during the last 30 million yr.
The Callitris clade is by far the most embolism-resistant group of trees in the world, containing multiple extreme P50 records
(i.e. C. glaucophylla (14.4 MPa), C. preissii (15 MPa), A. arenarius (15.2 MPa), C. columellaris (15.8 MPa), C. canescens 16.9 MPa) and C. tuberculata (18.8 MPa)). Only a handful of other gymnosperm species reach P50 values lower than
13 MPa, for example in Juniperus (Willson et al., 2008; Choat et al., 2012) and Tetraclinis (Bouche et al., 2014), and also conifers from the Cupressaceae family. Previous studies at broader evolutionary and geographical scales found a weak P50–
rainfall relationship, reporting both vulnerable and resistant species in areas of low precipitation (Brodribb & Hill, 1999;
Maherali et al., 2004; Choat et al., 2012; Brodribb et al., 2014). By contrast, our study at a regional, more restricted evolutionary scale reveals a strong relation between xylem embolism resistance and MAP (Fig. 2a). Further, our results suggest that variation in P50 can be largely predicted by local AI (Fig. 2b). These results
reveal two contrasting patterns in extant Callitris that either occupy high rainfall environments and have relatively vulnerable xylem or are distributed in xeric or seasonally dry environments with much less vulnerable xylem.
The poor fossil record of Callitris (Hill & Brodribb, 1999; Paull & Hill, 2010) and uncertainty in deep divergences in both our study and previous work (Pye et al., 2003; Piggin & Bruhl, 2010) shrouds in uncertainty the inference of ancestral ecology or drought tolerance in the Callitris clade. However, our mapping of P50across the Callitroidae suggests that the vascular apparatus
of the last common ancestor of Callitris was less resistant to drought-induced xylem embolism (Fig. 2a). In line with our results, the earliest definite fossil member of this clade is dated to c. 30 Ma and is associated with an assemblage of rainforest taxa, but shares characters with modern Callitris species found in both dry and wet environments (Paull & Hill, 2010). Before this time most of Australia was covered with warm and temperate rain-forests, with year-round wet climates, but the end of the Eocene (34 Ma) is marked by a global cooling event and a major drop in Ks (kg m –1 MPa –1 s –1 ) Dh (μm) P50 (MPa)
Tracheid frequency (n mm–2) Void-to-wood ratio (%)
Dh (μm) Tracheid frequency (n mm ) Void-to-wood ratio (%)
(a) (b) (d) (e) (f) 10 12 14 16 18 20 22 24 26 –20 –15 –10 –5 0 R² = 0.532 *** C_tu N_pa 1000 2000 3000 4000 1000 2000 3000 4000 R² = 0.414 *** C_in C_tu N_pa 20 25 30 35 40 (c) R² = −0.001 ns C_tu N_pa 0.50 0.60 0.70 0.80 R² = −0.016 ns C_tu N_pa R² = −0.022 ns C_in C_tuN_pa 10 12 14 16 18 20 22 24 26 C_in C_tu N_pa R² = −0.029 ns R² = 0.204 * C_in N_pa 20 25 30 35 40 R² = −0.043 ns C_in C_tu (g) Wood density (g cm ) 0.50 0.60 0.70 0.80 Wood density (g cm–3) (h) 0 0.5 1 1.5
C_in C_in C_in
N_pa
C_tu
Fig. 4 Relationships between wood construction traits and hydraulic safety and efficiency in Callitris. Correlations between embolism resistance (P50, upper panels) and xylem specific hydraulic conductivity (Ks, lower panels) and four wood traits: tracheid hydraulically weighted diameter (Dh) (a, e), tracheid frequency (b, f), void-to-wood ratio (c, g), and wood density (d, h). Linear regression lines are shown, with the adjusted R². Statistical significance is indicated by asterisks:***, P < 0.001; *, P < 0.05; ns, nonsignificance at a = 0.05. Error bars represent SE. The dotted line (a) is the best-fit nonlinear model. The shaded area in (d) shows trait space theoretically excluded owing to risk of tracheid implosion (Hacke et al., 2001). Red, aridity index (AI)< 0.5; green, 0.5< AI < 1; blue, AI > 1. C_tu, Callitris tuberculate; C_in, Callitris intratropica; N_pa, Neocallitropsis pancheri.
sea level (Martin, 2006; Macphail, 2007). This cooling trend is visible in the oxygen-isotope composition record (Fig. 6c) and is linked to lower rainfall based on paleosols from central Australia (Fig. 6b). Concomitantly, vegetation shifts indicate the existence of dry seasons and the emergence of the open woodland
ecosystems, at least in central Australia (Crisp et al., 2004; Martin, 2006; Fujioka & Chappell, 2010). The onset of more severe aridity in mainland Australia is dated to the mid-Miocene (Martin, 2006; Fujioka & Chappell, 2010; Metzger & Retallack, 2010). The extreme aridity over much of central Australia today is relatively young and linked to the glacial cycles in the northern hemisphere c. 1–4 Ma; (Crisp et al., 2004; Byrne et al., 2008). Over this period, highly resistant xylem seems to have evolved multiple times independently in different lineages within the group, as a result of convergent evolution. For instance, analo-gous low P50values are expressed in Actinostrobus and in the rest
of the Callitris clade. Meanwhile, our analysis found that the diversification of the Callitris clade happened in two successive pulses, the first at c. 30 Ma and the second from the mid-Miocene onwards (i.e. from c. 16 Ma to the present), coinciding with the major shifts in the climate of the region. This may underpin increasing aridification as a major driving force in Callitris evolutionary history, and the evolution of extreme xylem embolism resistance as a main trigger of their diversification, pro-viding Callitris with an outstanding competitive advantage over the last 30 million yr.
On the lower end of the Callitris xylem embolism resistance spectrum, we find a group of New Caledonian species (clade C; Fig 2a), which occur in tropical forests with significantly higher rainfall regimes. Previous work has shown that past climatic refu-gia may have persisted in the New Caledonia during the last glacial cycle, mitigating the significant drought that affected the region (Pouteau et al., 2015). The high P50values observed in the
New Caledonian clade may therefore result from a lack of drought-exerted selective pressure. Despite the remarkable diversity of conifers in New Caledonia (Jaffre et al., 1994; De Laubenfels, 1996), the Callitris clade is poorly represented with only three species, compared with 13 species of Araucaria. These are highly vulnerable to xylem embolism, with P50 typically in
the range of2 to 3 MPa (Bouche et al., 2014; Zimmer et al., 2016). Both these groups are threatened by climate change, as a result of their fragmented distributions and reduced habitat linked to human activity (Beaumont et al., 2011; Pouteau & Birnbaum, 2016). These taxa have different histories in Australia: Araucaria diversity has declined to only two extant species over the last 30 million yr (Kershaw & Wagstaff, 2001), probably because of their limited tolerance to drought (Zimmer et al., 2016). On the other hand, Callitris has thrived, largely thanks to the evolution of its xylem driven by reduced water availability. Coevolution of xylem traits
A drought tolerance strategy based on a highly embolism-resistant xylem is expected to come with costs, notably with higher carbon investment in conduit walls to cope with extreme xylem tensions and a reduction in water transport efficiency as a result of the associated reduction in conduit size (Hacke et al., 2001; Sperry, 2003; Pittermann et al., 2006a,b, 2011). High embolism resistance in xeric Callitris species is associated with small conduits, which is consistent with the trend in conifers and angiosperms (Sperry et al., 2006). Similarly to the trend across all
R² = 0.367 ** C_tu N_pa (b) Kth (kg m –1 MPa –1 s –1 ) 15 10 5 0 R² = −0.006 ns C_in C_tu N_pa (a) Ks (kg m –1 MPa –1 s –1) P50 (MPa) 1.5 1 0.5 0 –5 0 –15 –10 –20 4 2 0 R² = −0.047 ns (c) P50 (MPa) Kpit (kg m –1 MPa –1 s –1) C_in C_tu N_pa –5 0 –15 –10 –20 C_in
Fig. 5 Xylem safety vs efficiency relationship in Callitris. (a–c) Embolism resistance (P50) in relation to xylem specific hydraulic conductivity (Ks; a), theoretical conductance (Kth; b), and the derived total pit conductivity (Kpit; c). Linear regressions are reported with R². Statistical significance is indicated by asterisks:**, P < 0.01; ns, nonsignificance at a = 0.05. The dashed line (b) is the best-fit nonlinear model. Red, aridity index (AI)< 0.5; green, 0.5< AI < 1; blue, AI > 1. C_tu, Callitris tuberculate; C_in, C. intratropica; N_pa, Neocallitropsis pancheri.
conifers (Hacke et al., 2001), we found no species with high embolism resistance and low wood density. However, we found no evidence of an influence of aridity or P50 on wood density,
and the highest wood density was found in N. pancheri, the species with the highest P50measured in this study, a small
slow-growing tree found on ultramafic soils in hot tropical climates
with high rainfall. Wood density is a key functional trait that is determined by many factors, for example ring width, temperature and soil fertility (Chave et al., 2006; Heineman et al., 2016). It is worth noting here that all three New Caledonian species have high wood density and are found on skeletal, poor ultramafic soils, characterized by low concentrations of nitrogen, potassium,
Ice-fr ee temperatur e (°C) δ 18 O (‰ ) –1 0 1 2 3 4 5
Antarctic ice sheets
Northern hemisphere ice sheets
0 4 8 12
Partial or ephemeral Full-scale and permanent
?
Mid-Miocene Climatic optimum Mid-Eocene Climatic optimum Early-Eocene Climatic optimum ETM2 PETM (ETM1) 0 1 0 0 3 20 0 5 40 0 7 60 0 9 80 100 P. e n e c o i M Pli. Oligocene Eocene Paleocene Late CretaceousTime before present (million yr ago) 1 5 10 20 Log number of lineages 2 (a) (b) (c) 400 800 600 200 MAP (mm)
Fig. 6 Diversification of Callitris and climate trends of the past 65 million yr. (a) Lineage through time plot on log-scale. Two periods of higher diversification rate are indicated by black arrows. (b) Estimated mean annual precipitation (MAP; mm yr1) in Central Australia for the last 30 million yr, based on paleosol analysis (Metzger & Retallack, 2010). (c) Global temperature trend for the last 65 million yr. Data is deep-sea benthic foraminiferal oxygen isotopic composition (d18O;&, left axis), a proxy for global temperature (Zachos et al., 2001, 2008). The temperature scale (right axis) indicates the corresponding temperature, but is only valid for an ice-free world (i.e. not for periods with extensive ice at the poles (from c. 35 Ma)). Pli., Pliocene; P., Pleistocene. PETM, Paleocene-Eocene Thermal Maximum; ETM1-2, Eocene Thermal Maximum 1-2.
phosphorus, and calcium, and sometimes toxic concentrations of magnesium, nickel and manganese (Jaffre, 1995). Additionally, and in spite of a significant reduction in tracheid diameter, we found no significant decrease in xylem specific hydraulic conduc-tance with increasing aridity and increasing resisconduc-tance to embolism. This is in agreement with the lack of a linear trend across all conifers (Brodribb & Hill, 1999; Gleason et al., 2016).
How do embolism-resistant Callitris species maintain high water flow while reducing tracheid diameter? Xylem specific con-ductivity in conifers is strongly influenced by tracheid length and the total area of pits connecting adjacent tracheids (Pitter-mann et al., 2005). In Callitris, however, we found no evidence of a relationship between tracheid length or number of pits and Ks, while more resistant species did tend to have shorter tracheids
and fewer pits (Fig. S2). This could be linked to the increased probability of a rare ‘leaky’ pit in tracheids with more numerous pits, similar to the ‘rare pit’ hypothesis in angiosperms (Choat et al., 2003; Wheeler et al., 2005). Resistance of pits to water flow is mainly linked to the size and distribution of pores in the pit membrane and the depth and width of the pit aperture tun-nel (Hacke et al., 2004; Bouche, 2015). Embolism resistance is linked to the sealing of the pit aperture by the torus in conifers (Domec et al., 2006; Cochard et al., 2009; Delzon et al., 2010; Pittermann et al., 2010; Bouche et al., 2014). Increasing torus size relative to pit aperture (i.e. increasing torus overlap) has been shown to be related to more negative P50, but does not
reduce overall pit conductivity, which is mostly related to the size of pores in the margo and the pit aperture dimensions (Hacke et al., 2004; Pittermann et al., 2010; Bouche, 2015). Our results demonstrate that in the Callitris clade, water flow through vascular tissue is mostly limited by pit resistivity, indi-cated by Kpit<< Kth, as in other conifers (Pittermann et al.,
2006a). Additionally, bordered pits can be efficient for water transport, leading to high values of Kpit, and can also limit
air-seeding under high xylem tension (Pittermann et al., 2005; Domec et al., 2006). A prime example is C. intratropica, with average tracheid diameters (Dh= 14.2 lm; see Fig. 3) but by far
the highest xylem conductivity in Callitris (125 105m² MPa1s1), coupled with highly resistant xylem (P50= 12.8 MPa).
In other conifer genera, variation in xylem embolism resistance seems to be limited, for example in Pinus, in which P50 varies
between c.3 and 5 MPa (Delzon et al., 2010). The extraordi-nary lability of embolism resistance in the Callitris clade is proba-bly linked to their capacity to uncouple xylem safety from construction cost and xylem efficiency. Additionally, Callitris species display a typical anisohydric behavior, allowing the xylem midday water potential to track declining soil water potential dur-ing drought (Brodribb et al., 2013). Unlike many other conifer species, they use leaf desiccation rather than high concentrations of ABA to induce stomatal closure, allowing them to recover more quickly after rewatering (Brodribb & McAdam, 2013; Brodribb et al., 2014). However, continued leakage of water vapor from the leaf surface through closed stomata exposes their xylem to more negative water potential during prolonged drought – hence the need for more negative P50 to avoid embolism-induced loss of
hydraulic conductance. This derived strategy is followed by both Callitris and their sister-taxa in the Callitroidae subfamily (e.g. Austrocedrus, Widdringtonia, Fitzroya). A possible cost of this strategy could be the reinforcement of the leaves of these species, allowing them to endure extremely low leaf water potential before suffering leaf damage during drought (e.g. 8 MPa in Callitris rhomboidea (Brodribb et al., 2014), and 11 MPa in C. columellaris and C. preissii (Brodribb et al., 2010)). The more conservative water management strategy evident in many other conifers (e.g. Pinaceae, Araucariaceae) could help to explain the absence of very high embolism resistance in these clades (Delzon et al., 2010; Brodribb et al., 2014).
Our results confirm the evolutionary lability of xylem resistance to embolism, with a transition towards extreme resistance of several species within the Callitris clade. Combined with the onset of severe aridity in Australia, our results detail the remarkable adaptive radia-tion of this group. Despite overwhelming evidence for the role of decreasing water availability in shaping xylem traits, we found no evidence of its effect on xylem water transport efficiency and wood density. Species from more xeric areas develop smaller tracheids, but both wood traits and embolism resistance are totally discon-nected from overall xylem conductivity. This suggests a lack of tradeoffs between construction costs and safety, on the one hand, and efficiency on the other. We also highlight a likely evolutionary role of pit-level traits in this clade, regarding hydraulic safety and efficiency. While this clade has shown a remarkable capacity to adapt and even thrive in increasingly arid conditions, modern cli-mate change largely exceeds the pace of past clicli-mate upheavals. Many of the more vulnerable Callitris species have small restricted distributions and are already threatened by habitat destruction through land-use change and regular fires (IUCN, 2015), and could therefore be in severe danger of extinction in the coming decades.
Acknowledgements
The authors acknowledge Lin Kozlowski and Renee Smith for their analyses of wood anatomical traits, Anita Wesolowski and Vanessa Hequet for help with conducting sampling for this study, Frederic Lagane for assistance with the wood density measurements, and Gaelle Capdeville for assistance with embolism assessment. The authors thank Stacey Smith for helpful comments on a previous version of the manuscript, and three anonymous reviewers for their suggestions to improve the manuscript. Part of this study was funded by the programme ‘Investments for the Future’ (grant no. ANR-10-EQPX-16, XYLOFOREST) from the French National Agency for research, and mobility grants within the COTE Cluster of Excellence (grant no. ANR-10-LABX-45).
Author contributions
M.L., S.P., J-C.D. and S.D. designed the research; M.L., S.P., S.T. and S.D. collected the samples; N.N. provided DNA sequences; M.L. and S.P. measured the hydraulic traits; M.L. analyzed the data, ran the phylogenetic analysis, and led the manuscript preparation. All authors contributed to the writing of the manuscript.
References
Anderegg WRL, Berry JA, Smith DD, Sperry JS, Anderegg LDL, Field CB. 2012. The roles of hydraulic and carbon stress in a widespread climate-induced forest die-off. Proceedings of the National Academy of Sciences, USA 109: 233–237.
Anderegg WRL, Klein T, Bartlett M, Sack L, Pellegrini AFA, Choat B. 2016. Meta-analysis reveals that hydraulic traits explain cross-species patterns of drought-induced tree mortality across the globe. Proceedings of the National Academy of Sciences, USA 113: 2–7.
Antonelli AE, Verola CF, Parisod CN, Gustafsson ALS. 2010. Climate cooling promoted the expansion and radiation of a threatened group of South American orchids (Epidendroideae: Laeliinae). Biological Journal of the Linnean Society 100: 597–607.
Augusto L, Davies TJ, Delzon S, De Schrijver A. 2014. The enigma of the rise of angiosperms: can we untie the knot? Ecology Letters 17: 1326–1338.
Bailey IW. 1916. The structure of the bordered pits of conifers and its bearing upon the tension hypothesis of the ascent of sap in plants. Botanical Gazette 62: 133. Beaumont LJ, Pitman A, Perkins S, Zimmermann NE, Yoccoz NG. 2011.
Impacts of climate change on the world’s most exceptional ecoregions. Proceedings of the National Academy of Sciences, USA 108: 2306–2311. Benson DA, Karsch-Mizrachi I, Lipman DJ, Ostell J, Sayers EW. 2011.
GenBank. Nucleic Acids Research 39: D32–D37.
Booth TH. 2014. Using biodiversity databases to verify and improve descriptions of tree species climatic requirements. Forest Ecology and Management 315: 95–102. Bouche PS. 2015. Cavitation resistance and the functional role of bordered pits
in xylem of conifers– from inter-specific to within tree variability. Unpublished doctoral thesis, Universit€at Ulm, Germany.
Bouche PS, Jansen S, Sabalera JC, Cochard H, Burlett R, Delzon S. 2016. Low intra-tree variability in resistance to embolism in four Pinaceae species. Annals of Forest Science 73: 681–689.
Bouche PS, Larter M, Domec J-C, Burlett R, Gasson P, Jansen S, Delzon S. 2014. A broad survey of hydraulic and mechanical safety in the xylem of conifers. Journal of Experimental Botany 65: 4419–4431.
Bowman DMJS, Prior LD, Tng DYP, Hua Q, Brodribb TJ. 2011. Continental-scale climatic drivers of growth ring variability in an Australian conifer. Trees 25: 925–934.
Brodribb TJ, Bowman DMJS, Grierson PF, Murphy BP, Nichols S, Prior LD. 2013. Conservative water management in the widespread conifer genus Callitris. AoB Plants 5: 1–11.
Brodribb TJ, Bowman DJMS, Nichols S, Delzon S, Burlett R. 2010. Xylem function and growth rate interact to determine recovery rates after exposure to extreme water deficit. New Phytologist 188: 533–542.
Brodribb TJ, Cochard H. 2009. Hydraulic failure defines the recovery and point of death in water-stressed conifers. Plant Physiology 149: 575–584.
Brodribb T, Hill RS. 1999. The importance of xylem constraints in the distribution of conifer species. New Phytologist 143: 365–372.
Brodribb TJ, McAdam SAM. 2013. Abscisic acid mediates a divergence in the drought response of two conifers. Plant Physiology 162: 1370–1377. Brodribb TJ, McAdam SAM, Jordan GJ, Martins SCV. 2014. Conifer species
adapt to low-rainfall climates by following one of two divergent pathways. Proceedings of the National Academy of Sciences, USA 111: 14489–14493. Byrne M, Yeates DK, Joseph L, Kearney M, Bowler J, Williams MAJ, Cooper S,
Donnellan SC, Keogh JS, Leys Ret al. 2008. Birth of a biome: insights into the assembly and maintenance of the Australian arid zone biota. Molecular Ecology 17: 4398–4417.
Carpenter RJ, Hill RS, Jordan GJ. 1994. Cenozoic vegetation in Tasmania: macrofossil evidence. In: Hill RS, ed. History of the Australian
vegetation: cretaceous to recent. Cambridge, UK: Cambridge University Press, 276.
Chave J, Muller-Landau HC, Baker TR, Easdale TA, ter Steege H, Webb CO. 2006. Regional and phylogenetic variation of wood density across 2456 Neotropical tree species. Ecological Applications: A Publication of the Ecological Society of America 16: 2356–2367.
Choat B, Ball M, Luly J, Holtum J. 2003. Pit membrane porosity and water stress-induced cavitation in four co-existing dry rainforest tree species. Plant Physiology 131: 41–48.
Choat B, Cobb AR, Jansen S. 2008. Structure and function of bordered pits: new discoveries and impacts on whole-plant hydraulic function. New Phytologist 177: 608–626.
Choat B, Jansen S, Brodribb TJ, Cochard H, Delzon S, Bhaskar R, Bucci SJ, Feild TS, Gleason SM, Hacke UGet al. 2012. Global convergence in the vulnerability of forests to drought. Nature491: 752–755.
Cochard H, Damour G, Bodet C, Tharwat I, Poirier M, Ameglio T. 2005. Evaluation of a new centrifuge technique for rapid generation of xylem vulnerability curves. Physiologia Plantarum 124: 410–418.
Cochard H, H€oltt€a T, Herbette S, Delzon S, Mencuccini M. 2009. New insights into the mechanisms of water-stress-induced cavitation in conifers. Plant Physiology 151: 949–954.
Corlett RT, Westcott DA. 2013. Will plant movements keep up with climate change? Trends in Ecology & Evolution 28: 482–488.
Crisp M, Cook L, Steane D. 2004. Radiation of the Australian flora: what can comparisons of molecular phylogenies across multiple taxa tell us about the evolution of diversity in present-day communities? Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences 359: 1551–1571. Crisp MD, Cook LG. 2011. Cenozoic extinctions account for the low diversity of extant
gymnosperms compared with angiosperms. New Phytologist 192: 997–1009. David-Schwartz R, Paudel I, Mizrachi M, Shklar G, Delzon S, Cochard H,
Lukyanov V, Badel E, Capdeville G, Cohen S. 2016. Evidence for genetic differentiation in vulnerability to embolism in Pinus halepensis. Frontiers in Plant Science 7: 1–13.
Davis MB, Shaw RG. 2001. Range shifts and adaptive responses to Quaternary climate change. Science 292: 673–680.
De Laubenfels DJ. 1996. Gondwanan conifers on the Pacific Rim. In: Keast A, Miller S, eds. The origin and evolution of Pacific Island biotas, New Guinea to Eastern Polynesia: patterns and processes. Amsterdam, the Netherlands: SPB Academic Publishing, 261–265.
Delzon S, Cochard H. 2014. Recent advances in tree hydraulics highlight the ecological significance of the hydraulic safety margin. New Phytologist 203: 355–358. Delzon S, Douthe C, Sala A, Cochard H. 2010. Mechanism of water-stress induced
cavitation in conifers: bordered pit structure and function support the hypothesis of seal capillary-seeding. Plant, Cell and Environment 33: 2101–2111.
Dixon HH. 1914. Transpiration and the ascent of sap in plants. London, UK: MacMillan.
Domec J-C, Lachenbruch B, Meinzer FC. 2006. Bordered pit structure and function determine spatial patterns of air-seeding thresholds in xylem of douglas-fir (Pseudotsuga menziesii; Pinaceae) trees. American Journal of Botany 93: 1588–1600. Drummond AJ, Rambaut A. 2007. BEAST: Bayesian evolutionary analysis by
sampling trees. BMC Evolutionary Biology 7: 214.
Dupont LM, Linder HP, Rommerskirchen F, Schefuß E. 2011. Climate-driven rampant speciation of the Cape flora. Journal of Biogeography 38: 1059–1068. Eckenwalder JE. 2009. Conifers of the world: the complete reference. Portland OR,
USA and London, UK: Timber Press.
Eiserhardt WL, Borchsenius F, Plum CM, Ordonez A, Svenning JC. 2015. Climate-driven extinctions shape the phylogenetic structure of temperate tree floras. Ecology Letters 18: 263–272.
Farjon A. 2005. A monograph of Cupressaceae and Sciadopitys. Kew, UK: Royal Botanic Gardens, Kew.
Farjon A. 2010. A handbook of the world’s conifers (2 Vols.). Leiden, the Netherlands: BRILL.
Franklin GL. 1945. Preparation of thin sections of synthetic resins and woody resin composites and a new method for wood. Nature 155: 3924–3951. Fujioka T, Chappell J. 2010. History of Australian aridity: chronology in the
evolution of arid landscapes. Geological Society Special Publication 346: 121–139. Garamszegi Z. 2014. Modern phylogenetic comparative methods and their
application in evolutionary biology. London, UK: Springer.
Gleason SM, Westoby M, Jansen S, Choat B, Hacke UG, Pratt RB, Bhaskar R, Brodribb TJ, Bucci SJ, Mayr Set al. 2016. Weak tradeoff between xylem safety and xylem-specific hydraulic efficiency across the world’s woody plant species. New Phytologist 209: 123–136.
Guay R, Gagnon R, Morin H. 1992. A new automatic and interactive tree ring measurement system based on a line scan camera. The Forestry Chronicle 68: 138–141.
